Skip to main content



The project aims to understand the mechanism of how highly disease specific autoantibodies are generated in response to the exposure to a foreign antigen. IgA autoantibodies reactive with the enzyme transglutaminase 2 (TG2) are typical of coeliac disease (CD). These antibodies are only present in subjects who are HLA-DQ2 or -DQ8, and their production is dependent on dietary gluten exposure. This suggests that CD4+ gluten reactive T cells, which are found in CD patients and which recognise gluten peptides deamidated by TG2 in context of DQ2 or DQ8, are implicated in the generation of these autoantibodies. Using a novel technique, we have been able to visualise and isolate TG2-specific plasma cells from the small intestine of CD patients. On average 10% of the intestinal plasma cells in these patients are TG2 specific. From single TG2 specific plasma cells we have expression cloned immunoglobulin genes and produced monoclonal antibodies. A panel of nearly sixty TG2-specific monoclonal antibodies has been established. There is a strong bias for VH5-51 usage among the antibodies, and surprisingly the antibodies are modestly mutated. We have mapped the epitopes the antibodies recognise with a number of approaches including hydrogen-deuterium exchange experiments. The antibodies bind to a limited number of conformational epitopes which are clustered in the N-terminal part of TG2 and which are not displayed on the surface of cells. Key amino acids involved in the epitopes are mapped. Further, we have experimentally explored models as to how TG2 specific B cells can present gluten antigen to T cells. One model demonstrates that the enzyme TG2, upon binding, can crosslink gluten peptides onto a TG2-specific antibody or B cell receptor, hence facilitating uptake and presentation of the gluten peptide by the TG2-specific B cell. A second model demonstrates that TG2 is a very good substrate for itself so that the enzyme can create TG2 multimers by crosslinking. Gluten peptides can also be crosslinked into such TG2 multimers. These multivalent TG2 antigens decorated with gluten peptides bind to TG2 specific B cells so that gluten peptides are taken up and presented to gluten specific T cells. The two models are not mutually exclusive, and both of them place TG2 as a key factor in creating gluten T-cell epitopes. Finally, immunoglobulin(Ig) heavy and light chain knockin mouse strains expressing rearranged heavy chain VDJ genes and light chain VJ genes of one of the obtained anti-TG2 antibodies has been generated. The epitope recognised by the antibody is conserved between mouse and human TG2, and the double knock in mouse strain thus have B cells that recognise mouse TG2 as an autoantigen. B-cell tolerance to TG2 has been investigated by expressing the Ig knockin genes in mice which are TG2 proficient and deficient. We observe that the anti-TG2 BCR is expressed in a high number of B cells in periphery of the mice and there is no difference in expression of the BCR between proficient and deficient animals. Hence there is minimal or no B-cell tolerance induction in TG2 specific B cells. This is the most important conclusion of the project. This result suggests that the main control of antibody production to TG2 operates at the level of provision of T-cell help. Conceivably, gluten-specific T cells can provide such help by involvement the TG2-gluten complexes. Preliminary data indicate that gluten-specific T cells indeed can provide such help.